Qwen3.8 Max
NEWQwen · qwen/qwen3.8-max
Alibaba's flagship MoE with image and video input, hybrid thinking and a 1M-token context.
Not callable yet: Padma has not connected Qwen on this server. The price is real; requests answer 503 model_unavailable, and are not charged, until it is.
Input / 1M tokens
৳246.02
Output / 1M tokens
৳738.06
Cached input / 1M
৳30.75
Context length
1M
Max output
131.072K
Provider list price: $2.00 in · $6.00 out per 1M tokens (Qwen), × ৳123.01 per dollar. No markup. Prompt tokens the provider writes to its cache: ৳307.53 per 1M.
Suitable for
- Reasoning
- Coding
- Vision
- Agents
- Long context
Example request
curl https://padmarouter.com/v1/chat/completions \
-H "Authorization: Bearer $PADMA_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "qwen/qwen3.8-max",
"messages": [{"role": "user", "content": "Hello from Dhaka!"}]
}'Performance
Throughput40 tok/s
Time to first token2.6s
StreamingSupported
Tool callingSupported
Typical figures for this model, which the auto/* routers score on — not a guarantee for any one request.
Routing
Primary routeQwen direct
Fallback hostsNone — one host
Prompt retentionPadma keeps none unless you turn on payload logging; the provider's own API terms apply